Grasping the Keir A. Lieber Technological Peace

International Security Vol 25 No 1 (Summer 2000) pp 71ndash104copy 2000 by the President and Fellows of Harvard College and the Massachusetts Institute of Technology

71

Grasping theTechnological Peace

Keir A Lieber

The Offense-Defense Balance andInternational Security

Offense-defense the-ory argues that international conict and war are more likely when offensivemilitary operations have the advantage over defensive operations whereascooperation and peace are more likely when defense has the advantage Ac-cording to the theory the relative ease of attack and defensemdashthe offense-defense balancemdashis determined primarily by the prevailing state of technologyat any given time When technological change shifts the balance toward of-fense attackers are more likely to win quick and decisive victories Thisprospect of quick and decisive warfare exacerbates the security dilemmaamong states intensies arms races and makes wars of expansion preventionand preemption more likely When technological innovation strengthens thedefense relative to the offense states are more likely to feel secure and actbenignly1

Offense-defense theory has deep roots but has become increasingly popularin international relations scholarship and foreign policy analysis in recentyears The idea that the nature of technology affects the prospects for war and

Keir A Lieber is a Research Fellow in the Foreign Policy Studies Program at the Brookings Institution anda doctoral candidate in the Department of Political Science at the University of Chicago

For helpful comments and discussions I thank Stephen Biddle Meredith Bowers Jasen CastilloAlexander Downes David Edelstein Benjamin Frankel Charles Glaser Robert Lieber John Mear-sheimer Robert Pape Jordan Seng Stephen Walt Paul Yingling two anonymous reviewers andparticipants in the Program on International Politics Economics and Security at the University ofChicago I would also like to thank the John D and Catherine T MacArthur Foundation theAndrew Mellon Foundation and the Smith Richardson Foundation for their generous support ofthis research

1 The foundational works on offense-defense theory are Robert Jervis ldquoCooperation under theSecurity Dilemmardquo World Politics Vol 30 No 2 (January 1978) pp 167ndash214 George H QuesterOffense and Defense in the International System (New York John Wiley and Sons 1977) and StephenVan Evera Causes of War Power and the Roots of Conict (Ithaca NY Cornell University Press1999) especially chap 6 For crucial theoretical developments renements and extensions seeStephen Van Evera ldquoOffense Defense and the Causes of Warrdquo International Security Vol 22 No4 (Spring 1998) pp 5ndash43 Charles L Glaser and Chaim Kaufmann ldquoWhat Is the Offense-DefenseBalance and Can We Measure Itrdquo International Security Vol 22 No 4 (Spring 1998) pp 44ndash82Sean M Lynn-Jones ldquoOffense-Defense Theory and Its Criticsrdquo Security Studies Vol 4 No 4(Summer 1995) pp 660ndash691 and Charles L Glaser ldquoRealists As Optimists Cooperation As Self-Helprdquo International Security Vol 19 No 3 (Winter 199495) pp 50ndash90

peace is simple powerful and intuitively plausible Thus the offense-defensebalance concept has been used to address a variety of important historicaltheoretical and policy questions even when scholars have not adopted thebasic assumptions and logic of the theory2 Perhaps the most important reasonoffense-defense theory continues to appeal to scholars is that it offers a com-pelling argument for why intense security competition among states is not aninevitable consequence of the structure of the international system Specicallyfor realists who believe that threats are more important than raw materialpower in explaining state behavior the offense-defense balance appears toprovide a systematic method of predicting when the balance of power isthreatening and when it is not3

Offense-defense ideas also continue to shape contemporary foreign policydebates on arms control conventional and nuclear deterrence and force pos-ture the prevention of civil and ethnic conict and the so-called revolution inmilitary affairs On the latter issue for example the Economist recently pro-claimed that the world is in the early stages of a new military revolution thatwill strengthen the offense relative to the defense and thus create ldquoa strongincentive to strike rstrdquo4 The most policy-relevant conclusion offered by of-fense-defense theory is that arms races conict and war may be preventedthrough carefully designed arms control agreements that either deliberatelyshift the balance of technology toward defense or seek to correct mispercep-tions of the balance5

2 The offense-defense balance has been used to explain the origins of interstate war ethnic andcivil conict arms control arms racing alliance behavior military doctrine the consequences ofrevolutions grand strategy and the structure of the international system See discussion andreferences in Glaser and Kaufmann ldquoWhat Is the Offense-Defense Balancerdquo pp 44ndash45 nn 2 3and Lynn-Jones ldquoOffense-Defense Theory and Its Criticsrdquo pp 660ndash662 nn 3 43 For defensive realists the degree to which one state threatens another is a function of the relativedistribution of power (ie capabilities) ltered through the offense-defense balance If power isdistributed roughly equally and the balance of technology does not heavily favor the offense statesmay feel secure and signal their peaceful intentions If the balance heavily favors offense howeverstates will face strong incentives to build offensive forces and ght preemptively On the relation-ship between offense-defense theory and structural realism see Van Evera Causes of War pp 7ndash11117 255ndash256 Glaser and Kaufmann ldquoWhat Is the Offense-Defense Balancerdquo pp 48ndash49 Lynn-Jones ldquoOffense-Defense Theory and Its Criticsrdquo pp 660 n 1 664ndash665 and Glaser ldquoRealists AsOptimistsrdquo pp 54 60ndash644 ldquoSelect Enemy Deleterdquo Economist March 8 1997 pp 21ndash24 at p 215 See Jervis ldquoCooperation under the Security Dilemmardquo pp 199ndash201 Glaser and KaufmannldquoWhat Is the Offense-Defense Balancerdquo p 44 and Van Evera ldquoOffense Defense and the Causesof Warrdquo p 40 Other proponents of offense-defense theory are less sanguine about the capacity ofarms control to reduce the likelihood of war pointing out that arms control may be possible onlywhen it is not necessary

International Security 251 72

In this article I argue that the central concept of offense-defense theorymdashtheoffense-defense balance of technologymdashis deductively and empirically awedMy analysis conclusion follows from two basic questions First is there anoffense-defense balance of technology that can be used to predict militaryoutcomes Second do perceptions of the offense-defense balance affect politi-cal decisions to initiate conict I conclude that scholars have overstated boththe degree to which the balance of technology shapes battleeld outcomes andthe inuence that beliefs of offense or defense dominance have on political orstrategic behavior

The article is organized as follows First I examine the offense-defensebalance concept I present the basic denitions and assumptions required tooperationalize the balance for empirical evaluation and discuss why the ldquocorerdquoversion of the balance (which focuses solely on technology) is better than theldquobroadrdquo conception (which incorporates additional variables) Next I identifyand evaluate the most frequently employed criteria for classifying how tech-nology gives a relative advantage to offense or defense Finally I assess howoffense-defense explanations fare against the record of four watershed techno-logical innovations since 1850 These illustrative cases are the emergence ofrailroads in the nineteenth century the artillery and small arms revolution ofthe late nineteenth and early twentieth centuries the innovation of the tank inthe rst half of the twentieth century and the nuclear revolution of the latterhalf of the twentieth century The evidence suggests that although technologycan occasionally favor offense or defense perceptions of a technological bal-ance have little effect on the likelihood of war

The Offense-Defense Balance

The label ldquooffense-defense theoryrdquo refers to the body of work that exploreshow changes in the offense-defense balance shape state behavior in interna-tional politics Scholars working in the eld however have conceptualizedoperationalized and employed the balance differently6 This diversity notwith-standing there are several central features of the balance common to almostall approaches

6 See Sean M Lynn-Jones ldquoRealism Security and Offense-Defense Theories The Implications ofAlternative Denitions of the Offense-Defense Balancerdquo paper presented at the annual meetingof the American Political Science Association Boston Massachusetts September 3ndash6 1998

Grasping the Technological Peace 73

shared definitions and assumptionsThe offense-defense balance denotes some measure of the relative ease ofattacking and taking territory versus defending territory ldquoRelative easerdquo refersto the relative costs and benets of attacking versus defending The termsldquooffenserdquo and ldquodefenserdquo refer to actual military actions not the political inten-tions goals or objectives that motivate military action Specically offensemeans the use of military force to attack seize and hold a portion or all of adefender rsquos territory Defense involves using military force to prevent an at-tacker from seizing territory

The causal logic of offense-defense theory is based on the relative ease ofoffense and defense at the strategic level of war not the operational or tacticallevel The strategic level pertains to the highest levels of war planning anddirection and the achievement of ultimate war goals the operational level dealswith the conduct of specic campaigns within a theater of operations and thetactical level concerns actions taken within a particular battle7 The theoryultimately aims to explain decisions to initiate war what matters is leadersrsquoexpectations of nal war outcomes based on their perceptions of the strategicbalance Of course the feasibility of strategic offense and defense depends onthe success of operational and tactical offense and defense and thus the natureof warfare at these levels is highly relevant to understanding the overalloffense-defense balance

Proponents of the theory have struggled to offer a more precise denitionof the balance than just the relative ease of attack and defense8 Perhaps themost popular denition of the balance is cast in terms of a cost or investmentratio required for offensive success the ratio of the amount of resources thatan attacker must invest in offensive forces to offset the amount of resources adefender has invested in defensive forces9 Although measuring such a ratio(especially in historical cases) may be impossible for present purposes itstipulates that the offense-defense balance is a continuous not a dichotomous

7 Edward Luttwak and Stuart L Koehl The Dictionary of Modern War (New York Gramercy 1998)pp 568 442 5988 Jack S Levy has effectively highlighted the logical and methodological problems of severalprevious attempts to dene the balance in Levy rdquoThe OffensiveDefensive Balance of MilitaryTechnology A Theoretical and Historical Analysisldquo International Studies Quarterly Vol 28 No 2(June 1984) pp 222ndash2309 For variations on this denition see Jervis rdquoCooperation under the Security Dilemmaldquo p 188Glaser rdquoRealists As Optimistsldquo p 61 Lynn-Jones rdquoOffense-Defense Theory and Its Criticsldquo p 665Glaser and Kaufmann rdquoWhat Is the Offense-Defense Balanceldquo pp 3 7ndash10 and Robert GilpinWar and Change in World Politics (Cambridge Cambridge University Press 1981) p 62


variable What matters most for empirical evaluation is not whether the bal-ance in any given period favors offense or defense in absolute termsmdashin factit is almost always easier to defend than to attackmdashbut how and to what degreethe balance has shifted in either direction

Finally the offense-defense balance should be distinguished from two otherkey variables in international politics power and skill First the balance mustbe dened independently of the balance of power among states Battleeldoutcomes clearly depend on a host of factors other than the offense-defensebalance such as the relative distribution of military forces and resources Thusthe success of any given offensive or defensive strategy is not necessarilyindicative of the balance For example it might be misleading to say thatoffense is relatively easier than defense when a given state easily conquersanother state because this situation could simply have resulted from an over-whelming numerical disparity in the balance of forces rather than from anoffensive advantage Similarly a situation in which only one state in a conicthas acquired a new technology effectively represents a change in the balanceof power not in the offense-defense balance The effects of a new technologyon the balance can best be assessed in a conict between two roughly equal-sized military forces employing the technology Although the effects of theoffense-defense balance can theoretically overcome disparities in materialresources in determining war outcomes the two variables are analyticallydistinct

Second the offense-defense balance should be dened independently oflarge disparities in the level of skill between the attacker and the de-fender The relative ease of attack and defense is a concept meant to capturethe objective effects of military technology on war and politics To under-stand these baseline effects one must assume that states make reasonablyoptimal or rational decisions about force posture doctrine and strategyThe standard of optimality employed by proponents does not require thatattackers and defenders make the absolute best strategic choices whichmight be impossible to determine in any case Instead optimality in this con-text assumes that states make reasonably intelligent decisions about how toemploy existing technologies and forces given prevailing knowledge at thetime

core and broad versions of the balanceThe rdquocoreldquo version of offense-defense theory looks almost exclusively atchanges in military technology as the cause of shifts in the balance between


offense and defense10 Some scholars however incorporate a host of factors inaddition to technology when operationalizing the balance Proponents of therdquobroadldquo version include some or all of the following factors geography thecumulativity of resources (the ease of exploiting resources from conqueredterritories) nationalism regime popularity alliance behavior force size andmilitary doctrine posture and deployment11 For example according to pro-ponents of the broad approach nationalism tends to favor defense relative tooffense because people are more likely to ght harder when they believe theyare defending their rightful homeland from foreign invaders12 Incorporatinga host of geographic social political and military factors into the balanceclearly makes the theory more complex but proponents believe the coreversion is otherwise incomplete The relative ease of attack or defense isdetermined by a set of basic causal factors they argue and omitting thesefactors is unlikely to result in accurate explanations or predictions

Although proponents of the broad version of the offense-defense balancebelieve their approach strengthens offense-defense theory there are at leastthree important theoretical and practical advantages to focusing solely on thecore balance of technology First technology is the one determinant of thebalance common to all versions of the theory the most signicant factorshaping the balance and often the only factor analyzed in any detail byscholars None of the factors identied by the broad approach have such wideapplicability and importance Thus offense-defense theoryrsquos contribution to theconceptual toolbox of international relations largely turns on the role that thetechnological balance plays in shaping state behavior

Second because the broad offense-defense balance incorporates factorsunique to particular states (such as geography nationalism and regime popu-larity) it is not a systemic variable and the resulting theory is no longer

10 Adopting the core approach are George Quester Robert Jervis and Sean Lynn-Jones AlthoughJervis cites technology and geography as the two main factors that determine whether offense ordefense has the advantage technology is far more important in his work See Jervis rdquoCooperationunder the Security Dilemmaldquo pp 194ndash19611 Adopting the broad version are Stephen Van Evera Charles Glaser Chaim Kaufmann and tosome extent Ted Hopf and Jack Snyder For discussions of these factors see Van Evera rdquoOffenseDefense and the Causes of Warldquo pp 17 n 23 19ndash22 Glaser and Kaufmann rdquoWhat Is theOffense-Defense Balanceldquo pp 41 64ndash70 Lynn-Jones rdquoOffense-Defense Theory and Its Criticsldquop 669 Ted Hopf rdquoPolarity the Offense-Defense Balance and Warldquo American Political ScienceReview Vol 85 No 2 (June 1991) pp 477ndash478 Jack Snyder The Ideology of the Offensive MilitaryDecision Making and the Disasters of 1914 (Ithaca NY Cornell University Press 1984) pp 21ndash26and Jervis rdquoCooperation under the Security Dilemmaldquo p 19512 Glaser and Kaufmann rdquoWhat Is the Offense-Defense Balanceldquo pp 66ndash67 The authors alsonote exceptions under which nationalism favors offense


structural Offense-defense theory claims to share the appeal of other structuraltheories in international relations because it focuses on the war-causing effectsof a variable that is essentially exogenous to states Technology in principleprovides similar constraints and opportunities for all states in a given interna-tional system13 Thus comparable patterns of state behavior should arise undersimilar technological balances in history At best the broadly dened balancemight shed light on a specic military conict but its effects are not gener-alizable across space and time

Finally and most important the core version makes for a much more parsi-monious theory than the broad version All things being equal simpler theo-ries are both easier to measure (and test) and more intuitively appealing Firstconsider the issue of measurement Although measuring the balance of mili-tary technology is extremely complicated it is certainly more feasible thanmeasuring a balance that incorporates a host of complex ambiguous andsometimes cross-cutting variables Even if one were able to accurately assessthe impact of the broad factors on the relative ease of attack and defense onewould still have to weigh the relative importance of each factor and aggregatethem into a single value of the balance If the balance cannot be measured thetheory has little explanatory power and prescriptive utility

In addition a theory that uses few variables to explain a class of phenomenain the real world is more satisfying than a theory built on all possible causesAdopting the broad balance renders offense-defense theory atheoretical itbecomes a grab bag of variables employed in a purely post hoc descriptiveenterprise Although each of the factors identied in the broad version mayshape the relative ease of attack and defense a laundry-list explanation is notintuitively appealing The outbreak of violence in any single case results froman inevitably complex set of opportunities and constraints motives and goalsand decisions and actions A more interesting theoretical issue however iswhether there is something about military technology itself that affects thelikelihood of war and peace

How Does Technology Affect the Offense-Defense Balance

What are the criteria used to identify how technology gives a relative advan-tage to offense or defense at any given time Without such coding criteriawe have no theoretical guidance for judging which factors contribute dispro-

13 See Lynn-Jones rdquoOffense-Defense Theory and Its Criticsldquo p 668


portionately to offense or defense on the battleeld and thus cannot deter-mine the offense-defense balance Proponents have struggled to provideobjective and consistent criteria for distinguishing between offensive anddefensive technologies14 The few explicit discussions of differentiation crite-ria have often been supported by ambiguous arguments or contradictoryexamples15

Despite the high degree of confusion there appears to be at least someconsensus that mobility innovations favor offense whereas repower innova-tions favor defense Not all offense-defense proponents make these claims andthe large majority who do would not argue that these are concrete laws ofmilitary history Nevertheless the mobility and repower criteria are the mostuseful clearly articulated and frequently employed hypotheses

hypothesis 1 mobility-enhancing technologies favor offenseAlmost all proponents of offense-defense theory believe that new or improvedtechnologies that enhance mobility contribute relatively more to offense thandefense16 In military terms mobility is the ability of troops and equipment to

14 Critics of offense-defense theory often argue that it is impossible to determine how technologyaffects the balance because it is very difcult to categorize weapons as offensive or defensive SeeLevy rdquoOffensiveDefensive Balanceldquo pp 219ndash238 John J Mearsheimer Conventional Deterrence(Ithaca NY Cornell University Press 1983) pp 25ndash27 Samuel P Huntington rdquoUS DefenseStrategy The Strategic Innovations of the Reagan Yearsldquo in Joseph Kruzel ed American DefenseAnnual 1987ndash1988 (Lexington Mass Lexington Books 1987) pp 35ndash37 Jonathan ShimshonirdquoTechnology Military Advantage and World War I A Case for Military Entrepreneurshipldquo Inter-national Security Vol 15 No 3 (Winter 199091) pp 190ndash191 and Colin S Gray Weapons DonrsquotMake War Policy Strategy and Military Technology (Lawrence University Press of Kansas 1993)chap 2 Proponents claim that the logic of the theory depends not on the ability to classify weaponsas entirely offensive or defensive but on whether given weapons make offense or defense easierThe disagreement is primarily semantic and in any case proponents contend that in practiceoffensive and defensive weapons and force postures are generally distinguishable Lynn-JonesrdquoOffense-Defense Theory and Its Criticsldquo pp 674ndash677 and Glaser and Kaufmann rdquoWhat Is theOffense-Defense Balanceldquo pp 79ndash8015 Given the prominence and inuence of offense-defense theory the paucity of coding criteriais remarkable For example Jervis was pessimistic about the ability to dene in theory or identifyin practice the offense-defense variables that shape the severity of the security dilemma and it isdifcult to nd any coding criteria in his seminal piece on offense-defense theory Jervis rdquoCoop-eration under the Security Dilemmaldquo Similarly Van Evera states that rdquomilitary technology canfavor the aggressor or the defenderldquo but provides no criteria for deciding the issue anywhere inhis book Van Evera Causes of War p 160 Charles Glaser and Chaim Kaufmann provide the mostexplicit discussion in Glaser and Kaufmann rdquoWhat Is the Offense-Defense Balanceldquo For a reviewand critique of attempts to classify the technological characteristics of offense and defense seeKeir A Lieber rdquoOffense-Defense Theory and the Prospects for Peaceldquo PhD dissertation Univer-sity of Chicago forthcoming chap 216 Glaser and Kaufmann rdquoWhat Is the Offense-Defense Balanceldquo pp 61ndash63 and Quester Offenseand Defense p 3


move from one place to another There are essentially three types of mobilitystrategic operational and tactical

Strategic mobility is the ability to transport military forces from the home-land to a theater of operations or from one theater to another Offense-defensetheorists argue that greater strategic mobility allows the attacker to expedi-tiously transport and supply its forces far from its own borders thus negatingthe defender rsquos geographic advantage Operational mobility is the ability tomove forces within a theater According to proponents greater operationalmobility allows the attacker to concentrate forces quickly to achieve a numeri-cal advantage on a small portion of the front rapidly exploit weak points in adefender rsquos line or outank a defender rsquos position altogether Tactical mobilityis the ability to move forces on the battleeld in the face of enemy reOffense-defense proponents argue that greater tactical mobility reduces thenumber of casualties suffered by an attacker because these losses are partly afunction of the amount of time that forces are exposed to enemy re in anassault

There are several counterarguments to the mobility-favors-offense explana-tion First in terms of strategic mobility it is not clear why the ability totransport and supply forces far from the homeland gives an attacker an advan-tage over a defender who already has this capability Once an attack is underway in fact the defender depends more than the attacker on the ability toquickly move forces to that theater Moreover the impact of strategic mobilityappears indeterminate when the defender relies more heavily than the attackeron reinforcement from overseas territories and allies17 Second in terms ofoperational mobility the attacker depends more on the element of surprisethan on mobility to achieve a successful breakthrough of a defender rsquos frontline whereas the defender places a premium on mobility to reinforce threat-ened points in the front18 Unless a breakthrough penetration or envelopmentoccurs so rapidly that the defender never has a chance to react and counterat-tack the defender would also seem to prot more from mobility once anadvance penetration is under way Third greater tactical mobility may be moreadvantageous to the defender than the attacker in several ways Tactical mo-bility allows defenders to trade space for time through a series of tacticalwithdrawals to fortied positions where they can continue to re on attackingforces In addition greater offensive tactical mobility may actually increase

17 Glaser and Kaufmann rdquoWhat Is the Offense-Defense Balanceldquo p 63 n 5918 Mearsheimer Conventional Deterrence pp 25ndash26


attacker casualties as the greater speed of an assault often comes at the priceof reconnaissance protection and preparatory artillery re19 Finally tacticalmobility is advantageous for the defender because the defender often mustseize the tactical counteroffensive to avoid defeat

hypothesis 2 firepower-enhancing technologies favor defenseThe other plausible criterion for assessing the offense-defense impact of newmilitary technologies is the characteristic of repower Firepower is a measureof the destructive power of the weapons or array of weapons available to sidesin a conict Firepower consists of not only explosive power but also rangeaccuracy and rate of re

According to most proponents technological innovations that enhancerepower capability are disproportionately advantageous to the de-fense20 First repower allows the defender to threaten the attackerrsquos con-centration of forces before an attack An attacker typically needs a localadvantage of combat power to pierce the defender rsquos forward defensesNumerical superiority requires density but the greater density of forcesprovides more targets for defensive re and thus more attacker casualtiesSecond repower favors defense because it reduces the mobility (ieoffensive power) of the attacker In the face of greater defensive re an attackermust seek more armored protection cover concealment and dispersalmdashallof which slow the attackerrsquos advance Finally defensive repower forcesthe attacker to provide its own covering re in the advance which slowsthe attack because of added weight and time required to reposition coveringre

There are several reasons to believe however that repower is as crucial inthe attack as it is in defense The attacker relies heavily on suppressive orcovering re to neutralize or inhibit defender forces weapons or reconnais-sance Suppressive re by an attacker reduces the amount of re faced byadvancing forces and can pin down defender forces until they can be overrunand destroyed More important most successful offensives require preparatorybombardments before the attack Preparatory barrages can shatter the moraleof defenders destroy defensive positions and disrupt defender reinforcementsand communication Finally just as the defender uses repower to disrupt

19 See Stephen Biddle rdquoThe Determinants of Offensiveness and Defensiveness in ConventionalLand Warfareldquo PhD dissertation Harvard University 1992 pp 68ndash7720 The most clear-cut logic behind the repower hypothesis is found in Glaser and KaufmannrdquoWhat Is the Offense-Defense Balanceldquo p 64


attacker concentrations of forces before an attack the attacker depends onrepower to disperse defender forces into greater depth away from the front-line Because repower especially artillery does the greatest damage to forcesthat are grouped together dispersal is the wisest option When a defenderdisperses however the force-to-force ratio shifts in the attackerrsquos favor makingoffensive breakthroughs more likely

Mobility Firepower and International Security

The mobility and repower criteria whatever their logical weaknesses pro-vide a clear blueprint for case selection and empirical evaluation of offense-defense theory I consider the four biggest technological innovations inmobility and repower in modern history railroads the artillery and smallarms revolution tanks and nuclear weapons21 In each case I focus on twoquestions First what impact did the innovation have on military outcomesThe relevant issue is whether a mobility innovation shifted the offense-defensebalance toward offense resulting in more quick and decisive victories for theattacker or whether a repower innovation shifted the balance toward defenseresulting in longer indecisive battles of attrition Second what impact did theinnovation have on political outcomes Here we need to ask if and howdecisionmakers thought about these revolutionary innovations in offense-defense terms and whether these perceptions made war more likely Spe-cically did leaders believe that the attacker or the defender was privilegedby the innovation Were leaders more inclined to initiate war when theybelieved offense was favored

emergence of railroadsThe introduction of steam-powered railroads in the second half of the nine-teenth century perhaps marked the greatest revolutionary development inmilitary mobility since the wheel Armies were suddenly able to move andsustain huge forces across vast distances at up to ten times the speed ofmarching troops The rst practical locomotive appeared in 1825 railroadsspread rapidly across the European continent in the 1830s and 1840s and by

21 With the exception of the gunpowder revolution in the fteenth and early sixteenth centuriestechnological progress in warfare before the nineteenth century was more gradual and evolution-ary than revolutionary See Martin van Creveld Technology and War From 2000 BC to the Present(New York Free Press 1991) and Bernard Brodie and Fawn M Brodie From Crossbow to H-Bomb(Bloomington Indiana University Press 1973)


1850 all the major powers had conducted eld exercises in moving and sup-plying troops by rail22

military outcomes The advent of railroads coincided with several rela-tively short and decisive conicts between 1850 and 1871 and thus wouldappear to support the hypothesis linking mobility improvements with offen-sive advantages A closer look at the evidence however reveals that thesebattleeld outcomes resulted primarily from large asymmetries in power andskill rather than from the offensive advantages of railways Moreover al-though no wars occurred among the European great powers between 1871 and1914 World War I suggests that railroads if anything favored the strategicdefender

In 1850 in one of the earliest strategic uses of railroads Austria quicklymobilized and transported 75000 soldiers by rail to Bohemia forcing Prussiato back down in an escalating crisis Prussia had a small rail network and pooradministrative organization at the time however and bungled its own mobi-lization to the front23 The war between Austria and France in northern Italyin 1859 saw the large-scale use of railroads for strategic concentration opera-tional reinforcement and even tactical movement of troops The French wereable to deploy 120000 soldiers in eleven days and once in the theater ofconict use railroads to quickly and unexpectedly shift forces to defeat theAustrians In this case however the Austrians were guilty of incompetentpreparation mobilization and transportation and probably would have beendefeated by the French even in the absence of rail transport24

22 Some proponents of offense-defense theory view the railroad case as an exception to themobility-favors-offense prediction They argue that railway mobility is more useful for defendersbecause rail networks can be destroyed by retreating defenders more easily than they can beextended by advancing attackers Van Evera rdquoOffense Defense and the Causes of Warldquo pp 16ndash17Glaser and Kaufmann rdquoWhat Is the Offense-Defense Balanceldquo p 63 and Quester Offense andDefense chap 8 This remains an important test case however First according to the general logicemployed by proponents railroad mobility should favor offense at the strategic level where theattacker can more quickly concentrate forces at the front to surprise andor overwhelm thedefender Second our condence in the mobility-favors-offense hypothesis would be considerablydiminished by an empirical nding that railroads actually favored defenders on the whole giventhat railroads marked such a revolutionary improvement in mobility Finally even if militaryoutcomes demonstrate the defensive advantages of railroads we can still evaluate offense-defensetheory based on how leadersrsquo perceptions of the impact of railroads affected their behavior 23 Edwin A Pratt The Rise of Rail-Power in War and Conquest 1833ndash1914 (Philadelphia JBLippincott 1916) p 8 and Dennis E Showalter Railroads and Ries Soldiers Technology and theUnication of Germany (Hamden Conn Archon 1975) pp 37ndash3824 Pratt Rise of Rail-Power pp 9ndash13 John Westwood Railways at War (San Diego Calif Howell-North 1980) pp 14ndash16 Brodie and Brodie From Crossbow to H-Bomb p 149 and Michael HowardWar in European History (Oxford Oxford University Press 1976) pp 97ndash98


Prussiarsquos quick and decisive victories in the Wars of German Unicationmdashagainst Denmark (1864) Austria (1866) and France (1870ndash71)mdashhave com-monly been attributed to the offensive power of the Prussian railroads In factrailroads had much less impact on the conduct of these wars than did Prussiarsquossuperior doctrine organization and material power

Against Austria the Prussian attackers made extensive use of their railwaysto mobilize and transport an unprecedented 200000 troops to the theater ofoperations within three weeks But after initial deployments the Prussians raninto great trouble supplying and sustaining their offensive beyond the rail-heads Prussian forces quickly outran their supply convoys leaving food andfodder rotting at hopelessly congested railheads From the crossing of theAustrian border to the decisive battle of Koumlniggraumltz railways were irrelevantto the outcome of the war25

The Austrians were decisively defeated by Prussia because of crucial asym-metries in power and skill not because of the inherent offensive power ofrailroads The Austrians had only one major railroad line leading into thetheater of war while Prussia had ve Because of this superior railway net-work as well as the excellent planning of the Prussian general staff Prussiawas able to mobilize and deploy more battle-ready troops to the eld than theAustrians The rapid Austrian defeat was also facilitated by the technical edgeof the Prussian infantryrsquos breech-loading rie compared to the Austrian muz-zle-loader and Prussiarsquos superior doctrine of maneuvering forces to assumethe tactical defensive to take advantage of modern repower26

The Franco-Prussian War reveals a similar story though in this case theplanned French offensive was defeated quickly and decisively The Frenchmobilization and concentration of forces was utterly incompetent despite itsstrategically superior rail network and was the single most important causeof Francersquos defeat Although Prussiarsquos mobilization of roughly 400000 troopsby railroad was organized and efcient the concentration of their forces onthe French frontier was relatively inept These problems exposed the Prussiansto potential defeat by a better organized and capable defender27

25 Martin van Creveld Supplying War Logistics from Wallenstein to Patton (London CambridgeUniversity Press 1977) pp 83ndash85 and Pratt Rise of Rail-Power pp 104ndash105 Ironically the unex-pected Prussian freedom of movement created by the need to live off the land instead of dependingon supplies from railheads helped them win a decisive victory Westwood Railways at War p 5726 Showalter Railroads and Ries pp 14ndash15 59ndash68 and Larry H Addington The Patterns of Warsince the Eighteenth Century (Bloomington Indiana University Press 1994) pp 53ndash54 94ndash9727 Thomas J Adriance The Last Gaiter Button (New York Greenwood 1987) pp 47ndash54 and PrattRise of Rail-Power pp 110ndash115


After the initial deployment of forces railroads played virtually no impor-tant role in the Prussian offensive into France The extended railway lines werejammed with trafc and vulnerable to French attacks the railheads could notkeep up with the advancing forces and the Prussians faced enormous difcul-ties in getting supplies from the railheads to the front The heavy artilleryammunition and forces conveyed by railroads did make possible the siege andbombardment of Paris but this occurred well after the decisive mobile phaseof the campaign was over28 Other sharp disparities in military skill andcapability compounded Francersquos dismal performance including the Frencharmyrsquos awed command and staff system lack of reserves and unsuitabletactical doctrine of massed frontal attacks against the Prussian Krupp steelried breech-loading artillery

The American Civil War (1861ndash65) offers additional evidence that railroadmobility had not shifted the balance toward offense The control of railroadswas crucial for both Union and Confederate forces given the vast territorialscale of military operations At critical times both sides used railroads toconcentrate forces at strategic points to hold off enemy offensives29 AlthoughUnion forces also often relied on long rail lines for communications andsupplies as they advanced deep into the South railroads were on balancemore useful for the Confederates ghting on the strategic defense Out-numbered and outgunned the Confederates depended on rapidly concentrat-ing separated forces against key segments of the Union army Railroads pro-longed the Civil War and made it more difcult to ght quick and decisivecampaigns30

The impact of railroads in World War I clearly contradicts the mobility-favors-offense hypothesis By the time of the war all sides in the conict hada good understanding of railroad technology had adopted appropriate doc-trines for its use and had a generally equal level of skill in its employmentAt the outbreak of war railroads moved soldiers weapons and supplies at an

28 Van Creveld Supplying War pp 96 104 and Westwood Railways at War p 6629 See for example the Confederate use of railroads at First Bull Run and Chickamauga and theUnion use of railroads to reinforce troops on the Chattanooga front after Chickamauga GeorgeEdgar Turner Victory Rode the Rails The Strategic Place of the Railroads in the Civil War (LincolnUniversity of Nebraska Press 1992) chaps 7 21 Robert C Black III The Railroads of the Confederacy(Chapel Hill University of North Carolina Press 1998) pp 184ndash191 and Thomas Weber TheNorthern Railroads in the Civil War 1861ndash1865 (Bloomington Indiana University Press 1952) pp180ndash18130 Westwood Railways at War pp 17 29 and Christopher R Gabel Railroad Generalship Founda-tions of Civil War Strategy (Fort Leavenworth Kans US Army Command and General StaffCollege 1997)


unprecedented pace and scale The enhanced strategic mobility conferred bythe railroad did not translate however into quick and decisive battleeldoutcomes as was demonstrated by the French use of railroads to shift re-sources to halt the Schlieffen Plan the initial German offensive and the Ger-man use of railroads to shift forces from the western to eastern front to defeatthe Russian offensive31

political outcomes Railroads did not confer an advantage on the attackerDid political and military leaders believe and act as if they did32 In fact thehistorical record ips the standard offense-defense hypothesis on its headConventional wisdom between 1850 and 1871 (when wars were more frequent)held that railroads favored the defender while the dominant view after 1871(when wars were infrequent) held that railroads favored the attacker

By the mid-nineteenth century although some commentators warned thatthe building of railroads would only facilitate a foreign invasion of the home-land the prevailing military view was that railroads would favor the defenderby greatly improving the defender rsquos ability to shift troops to counter anythreatened sector of the frontier Theoretical writings most notably by theeconomist Friedrich List even surmised that the defensive advantages ofrailroads would bring perpetual peace to the European continent33

Military leaders in Prussia which was surrounded by potential enemieswere especially quick to see the defensive benets of railroads Helmuth vonMoltke chief of the general staff beginning in 1857 thought that Prussia wouldeventually be attacked and believed that the defensive mobility provided byan extensive network of railroads could counterbalance Prussiarsquos disadvantagein sheer number of forces34 After building such a comprehensive railwaynetwork and despite the perception that railroads favored the defender Prus-sia promptly provoked three wars in less than a decade waging some ofthe most decisive offensive campaigns in history In fact it was largely be-cause the railroad made the defense of Prussian territory easiermdashthat istroops could be deployed rapidly from the center to any front or redeployedfrom front to frontmdashthat Prussia was able to act aggressively toward itsneighbors

31 See Westwood Railways at War p 143 and van Creveld Supplying War pp 111ndash14032 Some offense-defense proponents claim that while actual offense dominance has been ratherrare rdquoperceived offense dominance is pervasive and it plays a major role in causing most warsrdquoVan Evera Causes of War p 18533 Showalter Railroads and Ries pp 18ndash35 and Westwood Railways at War pp 8ndash12 91 19734 Van Creveld Supplying War p 88 and Showalter Railroads and Ries pp 18 28 43 56


After the Franco-Prussian War every state in Europe quickly concluded thatrailroads favored the attacker and strove to copy Prussian institutions for theuse of railways35 All the European general staffs believed that quick anddecisive victory would come to the side that mobilized and concentrated itstroops the fastest and the railroads were thought to provide the key Despitethe new dominant view that railroads favored the attacker however no waroccurred on the continent for the next forty years

small arms and artillery revolutionA technical revolution occurred in the late nineteenth and early twentiethcenturies with the development of ried breech-loading small arms and artil-lery of unprecedented range accuracy and rate of re The combined effectwas an enormous increase in repower that armies could bring to bear on thebattleeld

military outcomes The repower revolution rendered massed frontal as-saults exceedingly difcult and many conicts were marked by costly battlesof attrition On balance therefore it is fair to say that the new technologiesshifted the offense-defense balance toward the defender

As early as the Crimean War (1854ndash56) ries showed the potential to behighly effective defensive weapons against attacking infantry But the Ameri-can Civil War Wars of German Unication Russo-Turkish War (1877ndash78)Anglo-Boer War (1899ndash1902) Russo-Japanese War (1904ndash05) and World War Iprovide the best evidence that defenders armed with modern ries machineguns and artillery had gained an enormous advantage against assaultinginfantry

The defensive advantage conferred by repower has often been exaggeratedhowever Two sets of evidence are notable First the tactical impasses createdby repower technologies did not necessarily translate into strategic or opera-tional deadlock The clearest example is the Prussian method of using strategicenvelopment and anking maneuvers to place forces where they could employtactical defensive repower against the enemy rear or ank This fusion of thestrategic offensive with the tactical defensive contributed to Prussiarsquos quickand decisive victories against Austria in 1866 and France in 1870ndash7136 In the

35 Westwood Railways at War p 197 Quester Offense and Defense pp 77ndash83 Howard War inEuropean History p 101 and van Creveld Technology and War p 15936 See Gunther E Rothenberg ldquoMoltke Schlieffen and the Doctrine of Strategic Envelopmentrdquoin Peter Paret ed Makers of Modern Strategy from Machiavelli to the Nuclear Age (Princeton NJPrinceton University Press 1986) pp 296ndash325


American Civil War Union forces were able to bring the war to an end morequickly when they learned to employ a strategic offensivetactical defensivedoctrine in pursuit of the Confederates37 In fact every major war between1861 and 1905 was ultimately decided by strategic offensive maneuvers38 InWorld War I the Schlieffen Plan was modeled on earlier Prussian victories andalmost succeededmdashrecall the French ldquomiracle on the Marnerdquo

A second reason to question the degree to which technology favored defensein this period was the success of the Germans and then the Allies in carryingout a series of offensive breakthroughs in the later stages of World War I usinginnovative infantry and combined arms tactics As far back as the AmericanCivil War infantry learned with some success to break up from waves ofattacking troops into small groups that alternated advancing with providingcovering re while others moved forward But it was the German armyrsquosdecision in 1917 to introduce new ldquoinltration tacticsrdquo that provided a realtactical solution to the stalemate of trench warfare These tactics called for abrief surprise artillery bombardment aimed at disrupting narrow weak pointsin the enemy line followed by the quick penetration by small independentgroups of storm troops who were to bypass points of strong resistance andadvance as far as possible The Germans employed inltration tactics withgreat success in late 1917 and especially in the spring of 1918 with the famousLudendorff offensives39 Despite no major changes in technology the Luden-dorff offensives achieved signicant and unprecedented breakthroughs fol-lowed by deep advances behind the Allied lines40 These offensives of courseultimately failed on the strategic level as the Germans lacked the transporta-tion and logistical capabilities necessary to follow up on their tactical successesbut the Allies adopted similar tactics in their own offensives until the end ofthe war

Massed frontal assaults in the face of modern repower were clearly not theoptimal method of attack but it is impossible to say whether warfare would

37 Trevor N Dupuy The Evolution of Weapons and Warfare (Indianapolis Bobbs-Merrill 1980)pp 201ndash20238 Van Creveld Technology and War p 17739 Timothy T Lupfer The Dynamics of Doctrine The Changes in German Tactical Doctrine during theFirst World War (Fort Leavenworth Kans US Army Command and General Staff College 1981)Dupuy Evolution of Weapons and Warfare pp 225ndash229 and Hew Strachan European Armies and theConduct of War (London Routledge 1983) pp 142ndash14940 Stephen Biddle analyzes the rst of the Ludendorff offensives as a test of offense-defensetheory in Biddle ldquoRecasting the Foundations of Offense-Defense Theoryrdquo paper presented at theannual meeting of the American Political Science Association Boston Massachusetts September3ndash6 1998


have looked drastically different had inltration tactics been introduced earlierin World War I At a minimum the German offensives suggest that the defen-sive advantage of repower was not as intrinsic to the prevailing technologyas is often portrayed

political outcomes In the decades before World War I according to of-fense-defense proponents European statesmen and military leaders errone-ously believed that attackers would benet most from the vast increases inrepower and wars would thus be short and decisive This ldquocult of theoffensiverdquo proponents argue was a principal cause of World War I41

Europeans did embrace offensive strategies before World War I but this hadlittle to do with beliefs in the offensive advantages of technology Instead thedominant preference for offensive strategies sprung from a host of organiza-tional social political and psychological causes Rather than address thesecauses of the cult of the offensive which have been well documented42 I focushere on whether perceptions of the nature of the repower revolution damp-ened or promoted conict

Prussian leaders were aware of the defensive impact of repower beforeinitiating the Wars of German Unication As early as 1858 Moltke arguedthat any potential enemy should be forced by maneuver into taking the tacti-cal offensive against Prussian defensive repower43 After the costly attacksby his forces against Denmark in 1864 Moltke concluded that in the age ofthe breech-loading rie no combination of bravery and superior num-bers could overcome the problem of attacking frontally over open groundagainst modern repower ldquoThe attack of a positionldquo Moltke wrote in 1865rdquois becoming notably more difcult than its defenserdquo44 This view of the

41 Van Evera Causes of War chap 7 Quester Offense and Defense chap 10 Jervis ldquoCooperationunder the Security Dilemmardquo pp 190ndash192 Stephen Van Evera ldquoThe Cult of the Offensive and theOrigins of the First World Warrdquo International Security Vol 9 No 1 (Summer 1984) pp 58ndash107 andSnyder Ideology of the Offensive For a critique of the ldquocult of the offensiverdquo argument see MarcTrachtenberg History and Strategy (Princeton NJ Princeton University Press 1991) chap 2 Therepower revolution and World War I is a crucial case for offense-defense theory James D Fearonand Richard K Betts note that the war has served as the principal source for generating offense-defense hypotheses and at the same time as the principal empirical test of these hypothesesFearon ldquoThe Offense-Defense Balance and War since 1648rdquo paper presented at the annual meetingof the International Studies Association Chicago Illinois February 21ndash25 1995 p 2 and BettsldquoMust War Find a Way A Review Essayrdquo International Security Vol 24 No 2 (Fall 1999) pp166ndash198 at p 184 Contradictory ndings would thus be especially problematic for the theory42 Van Evera ldquoCult of the Offensiverdquo and Snyder Ideology of the Offensive43 Strachan European Armies p 11544 Quoted in Showalter Railroads and Ries p 125 Prussian military instructions and trainingmanuals reected this belief as well


increased power of the defense did not dissuade Prussia from initiatingwar with Austria in 1866 and France in 1870 Instead Moltke adopted anoffensive strategy that sought to capitalize on the repower of a tactical defen-sive

Even after the spectacular offensive successes by Prussia no country deniedthe impact of repower45 Technical arguments that improvements in repowerhad beneted offense over defense did appear but were mainly promulgatedto justify offensive doctrines already deemed necessary for political and organ-izational reasons In short perceptions of the offensive advantages of repowerdid not lead states to adopt offensive strategies rather the bias in favor ofoffensive strategies made military thinkers concentrate on ways to use re-power in the attack

Germanyrsquos role in the outbreak of World War I contradicts offense-defensepredictions German military leaders evaluated the technical realities of re-power more objectively than all other European general staffs at the time andthus were fully aware of the increased power of the defense Yet Germanyrsquoswar plan since 1891 consistently called for a decisive offensive envelopmentagainst France before rapidly shifting forces against Russia Neither Alfred vonSchlieffen his successor Moltke (the younger) nor Germanyrsquos civilian leadersenvisioned that conquest would be easy Germanyrsquos geostrategic positioncombined with its foreign ambitions demanded a quick and decisive victoryat the outset of an expected two-front war46

Offense-defense predictions about European security between 1890 and 1914face other major anomalies Consider Stephen Van Everarsquos claim ldquoBelief inthe power of the offense increased sharply after 1890 and rose to very highlevels as 1914 approached [This belief] peaked in 1914 in Europe andGermany had the largest offensive opportunities and defensive vulnerabilitiesamong Europersquos powers Offense-defense theory therefore forecasts thatwar should erupt in Europe in about 1914 authored largely by Germanyrdquo47

Note rst that if war broke out in 1890 1905 or 1912 Van Evera could stillmake the same claim of theory validation because beliefs in offense weresteadily rising and thus potentially always ldquopeakingrdquo Second if offense was

45 According to one historian ldquonobody was under any illusion even in 1900 that frontal attackwould be anything but very difcult and that success could be purchased with anything short ofvery heavy casualtiesrdquo Michael Howard ldquoMen against Fire Expectations of War in 1914rdquo Inter-national Security Vol 9 No 1 (Summer 1984) p 4346 See Rothenberg ldquoMoltke Schlieffen and the Doctrine of Strategic Envelopmentrdquo47 Van Evera Causes of War pp 193 199 n 25


perceived to have had an advantage since 1890 (if not since 1871) why didwar not break out in 1890 1905 or 1912 Germanyrsquos alleged belief in thetechnical supremacy of the offensive combined with its sheer military ad-vantage over a weakened France and Russia ought to have led it to attackFrance in 1905 or Russia in 1909 when Germany faced real windows ofopportunity48 Moreover if the technological revolution in repower wasthought to favor offense why did no real competitive arms racing on landoccur before 1912 In short offense-defense theory does not appear capable ofexplaining the outbreak and timing of World War I

tanksThe character of land warfare was transformed by the mechanization andmotorization of armies from the end of World War I through World War II Themost important military innovation in this period was the tank The combina-tion of technological advances in the internal combustion engine armoredprotection and radio communication greatly increased operational mobility onthe battleeld

military outcomes Proponents of offense-defense theory believe thatthe incorporation of tanks into the European armed forces resulted in great-er offense dominance49 In World War I and the interwar period howevertanks had a negligible affect on operational outcomes In World War IIthe most relevant evidence does not show the offensive superiority of tankforces

In World War I tanks occasionally contributed to tactical breakthroughs andpenetrations but ultimately could not translate any tactical successes intooperational victories In fact the most spectacular breakthroughs of the warsuch as the 1918 Ludendorff offensives were made possible by new infantrytactics not tanks50 The wars fought between 1919 and 1939 were primarily

48 David G Herrmann persuasively argues that the mere existence of a window of opportunitywas not enough to prompt preemptive German strikes in the decade before 1914 Herrmann TheArming of Europe and the Making of the First World War (Princeton NJ Princeton University Press1996) See also Richard Ned Lebow ldquoWindows of Opportunity Do States Jump through ThemrdquoInternational Security Vol 9 No 1 (Summer 1984) pp 147ndash18649 Jervis ldquoCooperation under the Security Dilemmardquo p 197 Van Evera Causes of War pp 123162 Lynn-Jones ldquoOffense-Defense Theory and Its Criticsrdquo p 676 and Glaser and KaufmannldquoWhat Is the Offense-Defense Balancerdquo p 6350 On the role of tanks in World War I see Williamson Murray ldquoArmored Warfare The BritishFrench and German Experiencesrdquo in Murray and Allan R Millett eds Military Innovation in theInterwar Period (Cambridge Cambridge University Press 1996) pp 6ndash49 JP Harris Men Ideasand Tanks British Military Thought and Armoured Forces 1903ndash1939 (Manchester Manchester Uni-


either civil wars or colonial conicts and involved unevenly matched adver-saries or forces that were not well equipped with tanks When tanks were usedin battle as in the continuing French effort to extend control over Morocco(1908ndash34) the Italian war against Ethiopia (1935ndash36) the Spanish Civil War(1936ndash39) and the Russo-Japanese border clashes in Manchuria (1938ndash39) theresults were not illuminating51

The revolutionary potential of the tank was rst demonstrated by Germanyrsquosrapid envelopment of Polish forces in 1939 quick and decisive defeat of Francein 1940 and invasion of the Soviet Union in 1941 These campaigns offerinadequate evidence that tanks conferred a decisive advantage on the offensehowever because the victories resulted more from German material and doc-trinal superiority than from the balance of military technology In September1939 German forces were better trained better equipped and far larger thanthe Polish army52 The root cause of the German victory over France lies withGermanyrsquos far superior strategy tactics and organization rather than in thenature of its military hardware53 It is highly unlikely that Germany could haveachieved its stunning offensive in France if not for glaring Allied weaknessesand if the Allies had been more adept at using armored forces54 The Germaninvasion of the Soviet Union was enormously successful as panzer and mo-torized divisions achieved spectacular encirclements of entire Soviet armiesOnce again however the key to German success lay in Josef Stalinrsquos blundersand Red Army failings55

versity Press 1995) and Spencer C Tucker The Great War 1914ndash18 (Bloomington Indiana Univer-sity Press 1998)51 See Archer Jones The Art of War in the Western World (New York Oxford University Press1987) pp 497ndash507 and Addington Patterns of War pp 191ndash19452 The German armored and motorized divisions played an important role in overwhelming thePolish frontlines and encircling Polish forces but the German victory was a foregone conclusionone only hastened by Polish weaknesses and mistakes See Matthew Cooper The German Army1933ndash1945 Its Political and Military Failure (New York Stein and Day 1978) pp 169ndash176 BHLiddell Hart History of the Second World War (New York GP Putnamrsquos Sons 1971) pp 27ndash32 andJones Art of War pp 508ndash50953 The German armored forces were neither more numerous nor technically superior to Alliedforces In fact the Allies possessed a slight numerical advantage in tanks (at a ratio of 13 to 1)and manpower (12 to 1) Barry Watts and Williamson Murray ldquoMilitary Innovation in Peacetimerdquoin Murray and Millett Military Innovation in the Interwar Period pp 372ndash37354 For the ledger of German strengths and Allied weaknesses see Jones Art of War pp 510ndash544Larry H Addington The Blitzkrieg Era and the German General Staff 1865ndash1941 (New BrunswickNJ Rutgers University Press 1971) pp 101ndash123 Cooper The German Army pp 214ndash215 LiddellHart History of the Second World War chap 7 and Robert Allan Doughty The Breaking Point Sedanand the Fall of France 1940 (Hamden Conn Archon 1990) chap 155 On the Red Armyrsquos deciencies on the eve of the war see Alan Clark Barbarossa The Russian-German Conict 1941ndash45 (New York William Morrow 1965) chap 2


The best evidence with which to evaluate the offense-defense impact of tankscomes from military operations later in World War II when all sides in theconict had become relatively adept at armored warfare56 The most pertinentperiod is from operations on the eastern front in the winter of 1943 throughthe nal German offensive on the western front at the end of 1944 By thewinter of 1943 after two years of painful lessons the Russians had developedthe necessary organizational and doctrinal expertise to conduct offensive anddefensive armored warfare at a reasonably procient level After 1944 theimpact of tanks on the offense-defense balance was eclipsed by the sheerimbalance of material power between Germany and its adversaries

The evidence from this period demonstrates that the mobility conferred bytanks did not favor offense After the Russian encirclement of an entire Germanarmy at Stalingrad in late 1942 the Soviets went on the strategic offensive andthe Germans fell back on a fundamentally defensive strategy Time after timehowever the German army relied on its armored forces to halt and defeatmajor Soviet offensives German army doctrine regarded speed mobility andcounterattack to be the decisive elements of defense and tanks provided theperfect tool The Germans discovered early on that the best way to defeat aSoviet armored penetration was by immediate counterattack with tanksagainst the anks of the spearhead Thus as one panzer general noted ldquothearmored divisions originally organized as purely offensive formations hadbecome [by early 1943] the most effective in defensive operationsrdquo57

The power of a tank-oriented defense was best displayed by Field MarshalErich von Mansteinrsquos operations against major Soviet offensives in southernRussia and the Ukraine from January to March 1943 Though lacking the forcesnecessary to ght a true mobile defense Manstein allowed Soviet penetrationsin some sectors ordered stubborn positional defense in a few other sectorsand rapidly shifted and assembled panzer units for counterattacks against themost threatening breakthroughs Against a numerical balance of seven to oneManstein stabilized the southern front and prematurely ended the Sovietwinter offensives58 Mobile armored units proved uniquely suited for thesedelicate and demanding defensive operations

56 Agreeing are Glaser and Kaufmann ldquoWhat Is the Offense-Defense Balancerdquo p 5657 Quoted in Timothy A Wray Standing Fast German Defensive Doctrine on The Russian Frontduring World War II Prewar to March 1943 (Fort Leavenworth Kans US Army Command andGeneral Staff College 1986) p 17058 These operations are particularly informative because they were conducted largely free ofHitler rsquos rigid no-retreat policy which elsewhere prevented the German army from conducting a


The Germans were not alone in using the inherent mobility of tank forcesto stop enemy armored offensives The nal German strategic offensive of thewar on the eastern front aimed at pinching off and destroying the Soviet forcesin the Kursk salient in July 1943 This was a far more limited goal than thedeep penetrations and multiple encirclements of the earlier German offensivesbut Hitler intended to use the same successful formula of massing tanks for alightning blow against the Red Army59 The battle of Kursk the greatestarmored battle in history was indeed a quick one but resulted in a clear anddecisive victory for the defender The German armored spearheads were rap-idly worn down by antitank defenses and then crushed by the Soviet armoredreserves The defeat marked the rst time a German offensive had been haltedbefore it could break through enemy defenses into the strategic depths be-yond60

After Kursk Germany was pushed back along a broad front in an unrelent-ing series of Soviet offensives Despite facing a growing numerical imbalancein forces with a deteriorating army the Germans fought a skillful withdrawalshuttling dwindling armored reserves back and forth for effective counterat-tacks on Soviet armored breakthroughs61 The war ended on the eastern frontwith the Russians never having conducted any large strategic encirclementscomparable to the German victories of 1939 to 1941

The war on the western front provides generally less suitable evidence forexploring the offense-defense impact of tanks primarily because of the Alliedpreponderance of power62 In December 1944 Hitler launched his last majoroffensive of the war in the Ardennes forest The German armored spearheadspenetrated deep into the allied rear before they were halted by skilled armored

exible maneuver-oriented defense to which it was best suited and which would have been moreeffective against Soviet armored offensives For detailed accounts of these operations see DavidM Glantz From the Don to the Dnepr Soviet Offensive Operations December 1942ndashAugust 1943(London Frank Cass 1991) Erich von Manstein Lost Victories (Novato Calif Presidio Press 1982)FW Von Mellenthin Panzer Battles (New York Ballantine 1956) and Wray Standing Fast pp155ndash16459 Albert Seaton The Fall of Fortress Europe (New York Holmes and Meier 1981) p 5560 David M Glantz and Jonathan M House When Titans Clashed How the Red Army Stopped Hitler(Lawrence University Press of Kansas 1995) p 16761 For an account of these counterattacks particularly in November and December 1943 see ibidpp 174ndash17562 There are however some noteworthy instances of failed attempts to emulate the armoredoffensives of 1939 to 1941 For example Operation Cobra (the Allied breakout from Normandy inJuly and August 1944) saw armored breakthroughs and deep penetrations but the Allies wereunable to encircle the bulk of the German forces mainly because the tank forces had to operatein close conjunction with supporting infantry artillery and air forces to avoid destruction by othertanks and antitank forces


(and tactical air) maneuvers and counterattacks on the anks of the Germanbulge In sum the evidence suggests that the spectacularly successful armoredoffensives from 1939 to 1941 were an aberration not to be repeated againstopponents skilled at armored warfare

political outcomes Did a belief in the offensive superiority of tanks con-tribute to the outbreak of World War II Adolf Hitler may have been undeter-rable of course but we may still be able to assess whether his eagerness totake the offensive was inuenced by perceptions of the mobility-enhancingpotential of tanks Offense-defense proponents claim that Hitler was morewilling to attack his neighbors in 1939ndash41 because he believed offense wasdominant In particular proponents argue that Hitlerrsquos decision to attack theLow Countries and France in May 1940 (instead of in the fall of 1939 and winterof 1940) can be explained by his recognition that armored forces combinedwith the blitzkrieg doctrine had greatly strengthened the offense63

The interwar period witnessed a tremendous debate about how tanks shouldbe integrated into the armed forces but few experts concluded that tankswould have a revolutionary impact on warfare In the 1920s a small group ofBritish military thinkers led by Major-General JFC Fuller and Sir Basil LiddellHart argued that the inherent mobility of tanks could restore offensive supe-riority to the battleeld British experiments with armor never lived up toexpectations however and by the mid-1930s Fuller and Liddell Hart had losttheir enthusiasm for tanks as revolutionary offensive weapons64

From the end of World War I through the attack on Poland in September1939 few German leaders perceived the operational or strategic signicanceof tanks In the 1920s the German army had returned to its traditional militarydoctrine of seeking quick and decisive victories through highly mobile offen-sive warfare65 Most Germans nevertheless discounted the combat potentialof tanks based on their experience in World War I where tanks were seen tohave a ldquomoral effectrdquo on unprepared troops but could be easily defeated bycountermeasures66

63 Van Evera Causes of War pp 123 175 17764 On Fuller rsquos views see JFC Fuller The Reformation of War (New York EP Dutton 1923) pp152ndash169 Brian Holden Reid JFC Fuller Military Thinker (New York St Martinrsquos 1987) pp 57183 and chap 7 and Harris Men Ideas and Tanks chaps 6ndash8 On Liddell Hart see John JMearsheimer Liddell Hart and the Weight of History (Ithaca NY Cornell University Press 1988)pp 36ndash45 105ndash12365 See Cooper The German Army66 Wray Standing Fast p 6


A myth persists that during the 1930s the German army developed a newdoctrine of warfaremdashthe blitzkriegmdashbased on the revolutionary potential ofarmored forces to achieve a quick and decisive victory for the attacker whichthey then employed with great success in 1939ndash4167 It is true that HeinzGuderian a captain and then general in the German army was the drivingforce behind the development of tank forces and leading advocate of the ideaof using large independent armored formations to break through the enemyrsquosfront and conduct deep strategic penetrations Guderianrsquos ideas however metwith much skepticism resistance and outright subversion by the senior lead-ers of the army throughout the 1930s68 After all the German attack on Polandwas based not on a blitzkrieg strategy but on the traditional German strategyof a combined-arms attack on the anks in search of a decisive envelopmentof enemy forces Hitler eagerly attacked Poland absent any revelations of theoffensive power of tanks

Hitler was not deterred from attacking France in late 1939 and early 1940and planned to do so without any new model for employing tanks At theconclusion of the Polish campaign Hitler met with his military commandersand announced that he had decided to attack in the west as soon as possibleThe operational plan drawn up by the German army and endorsed by Hitlerin October 1939 called for an attack through the Netherlands Belgium andLuxembourg to defeat as much of the French and Allied forces as possible andto capture a large portion of the English Channel coast for subsequent opera-tions against Britain and the remainder of French territory German militaryleaders as well as Hitler himself realized that this plan could achieve only alimited territorial objective and would probably lead to a war of attrition Mostimportant Hitler did not believe that tank forces offered the potential for adecisive victory69

Despite Hitlerrsquos own reservations and the determined opposition of theGerman military leadership he gave his full approval to the plan to attack inthe west and xed November 12 as the date for the beginning of the offensiveThis attack was postponed because of poor weather conditions as were a series

67 For an excellent summary and critique of this view see JP Harris ldquoThe Myth of BlitzkriegrdquoWar in History Vol 2 No 3 (1995) pp 335ndash35268 Cooper The German Army pp 148ndash158 See also Heinz Guderian Panzer Leader (New York DaCapo 1996) and Barry R Posen The Sources of Military Doctrine France Britain and Germany betweenthe World Wars (Ithaca NY Cornell University Press 1984) pp 205ndash21969 Cooper The German Army pp 179ndash215 Doughty The Breaking Point pp 19ndash25 and Mearshe-imer Conventional Deterrence chap 4


of rescheduled offensives through December A nal date for the attack wasset for January 17 1940 but on January 10 a German plane carrying secretdocuments relating to the offensive was forced to land in Belgium thus com-promising German intentions The unlikely sequence of weather delays theplane crash and the onset of winter forced Hitler to postpone his attack untilthe spring of 1940 and made drastic operational changes more attractive Hitlereventually accepted an alternative plan of concentrating panzer forces for asurprise attack through the Ardennes70 In sum the evidence of Germanplanning and decisionmaking before World War II indicates that the politicaldecision to initiate military conict preceded any perceptions of the greatoffensive potential of tank technology

nuclear weaponsProponents of offense-defense theory argue that the nuclear revolutionstrongly shifted the offense-defense balance toward defense71 The case ofnuclear weapons is unique of course and the traditional concepts of offenseand defense do not translate easily from the conventional to the nuclear levelMoreover the coding of nuclear weapons as defense dominant ows from amore complicated logic than the enhanced-repower criterion These differ-ences notwithstanding offense-defense theory offers a valid explanation fornuclear defense dominance and yields concrete and testable predictions aboutthe political effects of nuclear weapons

military outcomes The impact of nuclear weapons on battleeld out-comes must be based on logical deduction not empirical evidence Accordingto offense-defense proponents when all sides in a conict possess a securesecond-strike nuclear capability (ie when no side can launch an attack thatis successful enough to prevent retaliation from the other) the defender hasan enormous advantage over the attacker This conclusion is counterintuitiveand requires clarication because under conditions of mutual assured destruc-tion (MAD) no side can defend against a nuclear attack strictly speaking

Offense-defense theory codes nuclear weapons as defense dominant becauseit is relatively easier and less costly for states to maintain a retaliatory capabil-

70 Some argue that even the attack launched in May 1940 was based on essentially traditionaloperational principles and methods and was not motivated by the belief that armored forces hadtransformed warfare See Harris ldquoThe Myth of Blitzkriegrdquo Cooper The German Army JP Harrisand FH Toase Armoured Warfare (New York St Martinrsquos 1990) pp 64ndash69 and Doughty TheBreaking Point p 32371 Robert Jervis The Meaning of the Nuclear Revolution Statecraft and the Prospect of Armageddon(Ithaca NY Cornell University Press 1989) chap 1 Charles L Glaser Analyzing Strategic NuclearPolicy (Princeton NJ Princeton University Press 1990) and Van Evera Causes of War chap 8


ity than to build a force capable of taking away anotherrsquos retaliatory capabilityStates are thus deterred from attacking one another in a nuclear world anddeterrence is the functional equivalent of defense The theory essentially aimsto explain when states feel secure and when they do not or alternatively whenthey can deter attacks and when they cannot When states rely on deterrencefor their security forces that enhance deterrence are essentially defensive In aworld of conventional arms deterrence becomes easier as the defender isincreasingly capable of denying territorial gains to the attacker In the nuclearworld deterrence rests on the defender rsquos ability to punish the attacker withunacceptable costs for attempted aggression The only way to take territory atan acceptable cost in the nuclear world is by eliminating the defender rsquos sec-ond-strike capability This is very difcult to do however because it is mucheasier to enhance onersquos own deterrent forces than to strengthen forces thatthreaten an adversaryrsquos deterrent forces72 Thus nuclear weapons favor thedefender by greatly improving the ability to deter by punishment

political outcomes The consequences of nuclear warfare in a MAD worldare easy to comprehend and extremely difcult to change Large and drasticshifts in the offense-defense balance such as has occurred with the nuclearrevolution should have signicant effects on international politics The mostfundamental prediction of offense-defense theory is that war among nuclearpowers should not occur The theory makes two additional predictions thatcan be evaluated against the historical record arms racing beyond robust MADlevels should not occur and security competition over distant territory shouldnot be intense

First consider the no-war prediction According to offense-defense theorythe prospect of devastation in a nuclear conict is enough to deter even themost highly expansionist country and the robust security provided by nuclearweapons virtually eliminates fears that might lead status quo states to launchpreventive or preemptive wars In short the implausibility of obtaining mili-tary victorymdashnot to mention a quick and decisive onemdashmakes war among thenuclear powers virtually obsolete

War among the major powers has not occurred since the introduction ofnuclear weapons Indeed it is hard to imagine any great power armed withthermonuclear weapons and advanced delivery systems being attacked andconquered in the traditional sense Alternative explanations for the ldquolong

72 Jervis ldquoCooperation under the Security Dilemmardquo p 198 Glaser Analyzing Strategic NuclearPolicy p 96 and Van Evera Causes of War pp 177ndash178 On deterrence by denial and deterrenceby punishment see Glenn H Snyder Deterrence and Defense Toward a Theory of National Security(Princeton NJ Princeton University Press 1961)


peacerdquo abound however specically that either bipolarity economic integra-tion or politicalnormative changes were responsible for preventing the ColdWar from becoming hot73 More important one could point to important casesthat cut against the logic of the no-war prediction For example a state armedwith nuclear weapons has been attacked (Israel in 1973) a state has intervenedin a war against a nuclear power (China in Korea against the United States in1950) and two states possessing nuclear weapons have fought each other (themajor armed clashes between China and the Soviet Union in 1969)74 Further-more the military conict between Pakistan and India two nuclear powersin the Kargil mountains of Kashmir in the spring of 1999 was a short war bytraditional casualty measures75 Nuclear weapons nevertheless have been amajor force for preventing war

Second according to offense-defense theory arms racing should not occuronce states believe they have acquired the capability for assured nuclearretaliation76 This prediction has both a quantitative and qualitative elementIn quantitative terms adversaries will not be too concerned with comparingthe relative size of their nuclear arsenals because even large shifts in relativeforce levels pose little threat to the ldquoweakerrdquo sidersquos ability to retaliate and inictunacceptable damage The actual size of the superpower arsenals in the ColdWar however far exceeded any reasonable estimate of the capabilities requiredfor assured destruction or deterrence By the early 1960s both US and Sovietleaders perceived that the United States could not effectively disarm the Sovi-ets with a rst strike77 In terms of US requirements for deterrence theminimum force levels thought necessary to inict unacceptable damage on theSoviet Union in a retaliatory strike were well in hand by 196478 By this time

73 See John Lewis Gaddis The Long Peace Inquiries into the History of the Cold War (New YorkOxford University Press 1989)74 The severity of the 1969 Sino-Soviet conict has been underappreciated See Raymond LGarthoff Detente and Confrontation American-Soviet Relations from Nixon to Reagan rev ed (Wash-ington DC Brookings 1994) pp 228ndash242 and Henry Kissinger White House Years (Boston LittleBrown 1979) pp 183ndash19475 The best estimates are roughly 1200 killed in the ten-week conict The Indian army and airforce suffered 474 killed Government of India Report of the Kargil Review Committee (New DelhiGovernment of India March 2000) Executive Summary p 10 An estimate of regular Pakistaniarmy casualties is 700 killed Report of the Kargil Review Committee p 7576 See Jervis ldquoCooperation under the Security Dilemmardquo pp 188 198 and Van Evera Causes ofWar pp 244ndash24577 David Alan Rosenberg ldquoThe Origins of Overkill Nuclear Weapons and American Strategy1945ndash1960rdquo International Security Vol 7 No 4 (Spring 1983) pp 38ndash44 and David Holloway TheSoviet Union and the Arms Race (New Haven Conn Yale University Press 1983) chap 378 Most analysts cite Secretary of Defense Robert McNamararsquos famous criteria for an assureddestruction capability developed in 1963 which assumed the United States would need about


the United States had deployed 4718 strategic nuclear warheads on a triad ofbombers intercontinental ballistic missiles (ICBMs) and submarine-launchedballistic missiles (SLBMs) while the Soviet Union had deployed almost 800warheads Yet the Soviet Union had begun a massive military buildup of bothnuclear and conventional forces and the United States quickly followed witha vast increase of its own warheads The United States deployed 6135 deliv-erable warheads in 1970 10768 in 1980 12304 in 1990 and almost 7000 in1999mdasha decade after the end of the Cold War The Soviet Union deployed 2327warheads in 1970 7488 in 1980 11252 in 1990 and Russia still possessedalmost 5500 warheads in 199979

The qualitative aspect of the no-arms-race prediction is that once statesnd themselves in a MAD world they should not attempt to gain an advan-tage at the nuclear level by building offensive counterforce weapons (ieforces aimed at destroying an adversaryrsquos strategic nuclear weapons)Possession of an assured destruction capability already provides states witha high degree of security a rst-strike advantage is virtually unattainableimpossible to maintain and thus irrational to pursue The evolution of USand Soviet nuclear strategies from the early 1960s was characterized by apersistent interest in escaping from MAD through the deployment of sophis-ticated counterforce weapons systems Both countries threatened each otherrsquosretaliatory capabilities aimed at limiting damage to their own forces andsociety and generally sought to prevail in the event of nuclear war In theUS case policymakers declared a nuclear doctrine consistent with MADbut actually embraced a counterforce posture and strategy80 US counter-force planning and targeting began in the late 1950s even though it wasapparent that the Soviet Union would soon acquire a secure second-strike

1500 warheads For discussions of McNamararsquos criteria see Alain C Enthoven and K WayneSmith How Much Is Enough Shaping the Defense Program 1961ndash1969 (New York Harper and Row1971) pp 175 207ndash210 Fred Kaplan The Wizards of Armageddon (Stanford Calif Stanford Univer-sity Press 1983) chap 22 and Scott D Sagan Moving Targets Nuclear Strategy and National Security(Princeton NJ Princeton University Press 1989) pp 32ndash3479 All gures are from Robert S Norris and Thomas B Cochran US and USSRRussian StrategicOffensive Nuclear Forces 1945ndash1966 (Washington DC National Resources Defense Council 1997)and John Pike ldquoNuclear Forces Guiderdquo Federation of American Scientistshttpwwwfasorgnukeguidesummaryhtm (data as of January 1999)80 See Aaron L Friedberg ldquoThe Evolution of US Strategic Doctrine 1945ndash1980rdquo in Samuel PHuntington ed The Strategic Imperative New Policies for American Security (Cambridge MassBallinger 1982) pp 53ndash99 Desmond Ball ldquoThe Development of the SIOP 1960ndash1983rdquo in Ball andJeffrey Richelson eds Strategic Nuclear Targeting (Ithaca NY Cornell University Press 1986)pp 57ndash83 Sagan Moving Targets and Eric Mlyn The State Society and Limited Nuclear War (AlbanyState University of New York Press 1995)


capability81 This strategy was accelerated in the 1960s when more than 90percent of Soviet bloc targets in the US nuclear war plan were counterforcetargets82 In the 1970s despite having concluded agreements with the SovietUnion to limit strategic defense the United States continued to enhance itscounterforce arsenal by building highly accurate weapons capable of destroy-ing hardened Soviet targets83 In the 1980s the Reagan administration tookcounterforce to an extreme by pursuing effective strategic defenses and offen-sive counterforce programs84 The Soviet Union also did not regard the pos-session of an assured destruction capability as sufcient to maintain itssecurity The Soviets began to build an antiballistic missile (ABM) systemaround Moscow in the mid-1960s and after signing the ABM treaty in 1972continued to develop its strategic defense capabilities through air defenseprograms against bombers and civil defense efforts aimed at protecting Sovietleadership More important the Soviets embarked on a massive nuclear coun-terforce buildup that stressed heavy accurate missiles specically designed todestroy the US ICBM force85 Both the United States and the Soviet Unionrecognized their mutual vulnerability to nuclear destruction but were drivenby the goal of winning a nuclear war and accordingly based their nuclearstrategies on robust counterforce arsenals

A nal prediction offense-defense theory makes about behavior under nu-clear defense dominance is that states should not compete or ght too intenselyover territory beyond the homeland or the homeland of close allies Nuclear

81 Rosenberg ldquoThe Origins of Overkillrdquo Friedberg ldquoEvolution of US Strategic Doctrinerdquo andRobert Jervis The Illogic of American Nuclear Strategy (Ithaca NY Cornell University Press 1984)p 4482 Ball ldquoDevelopment of the SIOPrdquo pp 66ndash6783 Among other steps the United States built Minuteman ICBMs and Polaris SLBMs addedthreatening multiple independently targetable reentry vehicles (MIRVs) to its ICBMs and SLBMsand decided to upgrade its Minuteman III ICBMs and deploy highly lethal and accuratePeacekeeper MX ICBMs Trident D-5 SLBMs and Pershing II medium-range ballistic missilesmdashallof which threatened the Soviet ability to retaliate in a nuclear exchange84 Strategic defense efforts fell under the Strategic Defense Initiative (SDI) and a host of air-defense early-warning and civil defense programs Offensive programs included accelerating theTrident D-5 program and building new bombers cruise missiles warheads and sensors Barry RPosen and Stephen Van Evera ldquoDefense Policy and the Reagan Administration Departure fromContainmentrdquo International Security Vol 8 No 1 (Summer 1983) pp 3ndash45 and Desmond Ball andRobert C Toth ldquoRevising the SIOP Taking War-Fighting to Dangerous Extremesrdquo InternationalSecurity Vol 14 No 4 (Spring 1990) pp 65ndash9285 See Robert P Berman and John C Baker Soviet Strategic Forces Requirements and Responses(Washington DC Brookings 1982) chap 3 Holloway Soviet Union and the Arms Race chap 3William T Lee ldquoSoviet Nuclear Targeting Strategyrdquo in Ball and Richelson Strategic Nuclear Target-ing and David Miller The Cold War A Military History (New York St Martinrsquos 1998) pp 98ndash102


weapons devalue traditional concerns over geographic depth in other wordsbuffer zones and distant bases are less important in a nuclear world becausenuclear retaliation can be assured in their absence86 In the Cold War offense-defense theory predicts minimal intervention and competition between thesuperpowers in the third world ldquoNuclear weapons make conquest muchharder and vastly enhance the self-defense capabilities of the superpowersThis should allow the superpowers to take a more relaxed attitude towardevents in third areas including the Third World since it now requires muchmore cataclysmic events to shake their defensive capabilities Whatever hadbeen the strategic importance of the Third World in a nonnuclear worldnuclear weapons have vastly reduced itrdquo87

Although a full exploration of the nature of US and Soviet interventionand competition in the third world is beyond the scope of this article it isfair to say that the superpowers had anything but a ldquorelaxed attituderdquo Infact most crises between the United States and the Soviet Union occurredin the third world as each superpower resorted to the whole range of eco-nomic political and military means to advance its own interests or blockthe inuence of its rival88 The most relevant and striking evidence is therelationship between the Soviet achievement of nuclear parity and its in-creased level of intervention in the third world As several scholars havenoted the Soviets were constrained from too overtly challenging the UnitedStates in the third world early in the Cold War because of US nuclear hegem-ony The arrival of strategic nuclear parity however coincided with a muchmore assertive role for the Soviets as demonstrated by their actions in theMiddle East (1970ndash73) Angola (1975ndash76) Ethiopia (1977ndash78) Yemen (1978ndash79)and Afghanistan (1979)89 The emergence of nuclear parity did not mitigateand may have aggravated the superpower competition for inuence in thethird world

The United States intervened with its own military forces in the third worldthroughout the Cold War as well including in Korea (1950ndash53) Egypt (1956)

86 Van Evera Causes of War p 24587 Posen and Van Evera ldquoDefense Policy and the Reagan Administrationrdquo p 3388 See Garthoff Detente and Confrontation pp 732ndash74589 See Bruce D Porter ldquoWashington Moscow and Third World Conict in the 1980srdquo in Hunt-ington Strategic Imperative pp 258ndash259 and Coit Blacker ldquoThe Kremlin and Detente SovietConceptions Hopes and Expectationsrdquo in Alexander George ed Managing US-Soviet Rivalry(Boulder Colo Westview 1983) pp 127ndash128


Lebanon (1958 and 1982) Thailand (1962) Laos (1962ndash75) Vietnam (1964ndash73)Congo (1964 and 1967) the Dominican Republic (1965) Cambodia (1970)Libya (1981 and 1986) Grenada (1983) and Panama (1989)90 Because greatpowers have interests beyond defense of the homeland that may requireintervention and competition abroad a superpower military presence andsome competition for power in the third world do not necessarily undermineoffense-defense theory The fact that the superpowers were consistently drawninto very real and frequently costly conicts in areas of little economic orstrategic value however is disconrming evidence

Conclusions

Offense-defense theory contends that the relative ease of attack and defenseoften plays a major role in causing instability and war in international politicsThe theory holds that states will tend to seek security through aggression whenoffensive advantages render the capabilities of others more threateningwhereas peace is more likely when defensive advantages make changes in thebalance of military power less worrisome The theory also claims prescriptiveutility because misperceptions and miscalculations of the balance often leadstates to initiate conict when they otherwise might feel secure with the statusquo these misperceptions can sometimes be ameliorated through arms controland condence-building measures

Given the pervasive inuence of the offense-defense balance in internationalsecurity scholarship the logical consistency and empirical validity of the the-ory deserve rigorous evaluation The lionrsquos share of past criticism has beendirected at the conceptual and operational problems endemic to ldquobroadrdquo ver-sions of offense-defense theorymdashthose versions that dene the balance toinclude a laundry list of factors in addition to technology Although proponentsof the broad theory believe that they are buttressing the explanatory power ofthe theory with a more sophisticated approach in practice the balance becomesan ad hoc collection of variables employed pell-mell to account for empiricalanomalies and logical qualications

The ldquocorerdquo version of offense-defense theory which focuses almost exclu-sively on how technology shapes the relative ease of attack and defense is thepotentially more fruitful approach The core theory offers two basic criteria for

90 Garthoff Detente and Confrontation p 744 and Ellen C Collier Instances of Use of United StatesForces Abroad 1798ndash1993 (Washington DC Congressional Research Service 1993)


judging how a given technology affects the offense-defense balance and thusmilitary outcomes mobility-improving innovations generally favor offenseand result in more quick and decisive victories for the attacker whereasrepower-enhancing innovations typically strengthen defense and lead tomore indecisive warfare In terms of political outcomes the theory predictsthat states are more likely to initiate conict when they perceive that theoffense-defense balance favors offense

This article ldquotestedrdquo offense-defense propositions using four illustrative casestudies chosen as the most important mobility and repower innovations inmodern military history The impact of railroads on military outcomes wasmixed with attackers beneting from greater strategic mobility and defendersproting from better operational mobility The quick and decisive character ofsome wars at the time primarily arose because of great asymmetries in militarystrength and doctrine More important more wars were initiated when rail-roads were perceived to favor the defender than when railroads were thoughtto favor the attacker The small arms and artillery revolution shifted theoffense-defense balance toward defenders Perceptions of the military conse-quences of the revolution in repower technology however did little to damp-en conict In fact leaders in Prussia and Germany were more cognizant ofthe technical realities of repower than those in any other European state andyet provoked all of the major power wars on the continent at the time Tankshad an indeterminate effect on the offense-defense balance The conventionalview that tanks favored offense is based largely on Germanyrsquos stunning op-erations from 1939 to 1941 and is undermined by the evidence when all sideswere adept at armored warfare Moreover Hitler attacked his neighbors absentany belief in the great offensive potential of tanks Finally the nuclear revolu-tion offers only mixed evidence for offense-defense theory The logic for codingnuclear weapons as defense-dominant is sound and nuclear war has notoccurred Although leaders correctly perceived the military consequences of anuclear conict however the United States and the Soviet Union engaged inan intense and costly arms race and competed hard in regions of secondaryimportance

The evidence suggests that scholars have overstated both the degree towhich the nature of technology shapes military outcomes and the inuencethat beliefs of offense or defense dominance have on political and strategicdecisions Understanding the limitations of offense-defense theory might helpscholars develop more nuanced causal explanations as well as construct moreprecise empirical tests of these hypotheses Ultimately however the theory


may not provide enough analytical leverage for understanding internationalpolitics especially given the complexity of operationalizing and measuring theoffense-defense balance In any event the relationship between technologicalchange and international security is too important and fascinating a subject toabandon based on the awed concept of an offense-defense balance of militarytechnology


peace is simple powerful and intuitively plausible Thus the offense-defensebalance concept has been used to address a variety of important historicaltheoretical and policy questions even when scholars have not adopted thebasic assumptions and logic of the theory2 Perhaps the most important reasonoffense-defense theory continues to appeal to scholars is that it offers a com-pelling argument for why intense security competition among states is not aninevitable consequence of the structure of the international system Specicallyfor realists who believe that threats are more important than raw materialpower in explaining state behavior the offense-defense balance appears toprovide a systematic method of predicting when the balance of power isthreatening and when it is not3

Offense-defense ideas also continue to shape contemporary foreign policydebates on arms control conventional and nuclear deterrence and force pos-ture the prevention of civil and ethnic conict and the so-called revolution inmilitary affairs On the latter issue for example the Economist recently pro-claimed that the world is in the early stages of a new military revolution thatwill strengthen the offense relative to the defense and thus create ldquoa strongincentive to strike rstrdquo4 The most policy-relevant conclusion offered by of-fense-defense theory is that arms races conict and war may be preventedthrough carefully designed arms control agreements that either deliberatelyshift the balance of technology toward defense or seek to correct mispercep-tions of the balance5

2 The offense-defense balance has been used to explain the origins of interstate war ethnic andcivil conict arms control arms racing alliance behavior military doctrine the consequences ofrevolutions grand strategy and the structure of the international system See discussion andreferences in Glaser and Kaufmann ldquoWhat Is the Offense-Defense Balancerdquo pp 44ndash45 nn 2 3and Lynn-Jones ldquoOffense-Defense Theory and Its Criticsrdquo pp 660ndash662 nn 3 43 For defensive realists the degree to which one state threatens another is a function of the relativedistribution of power (ie capabilities) ltered through the offense-defense balance If power isdistributed roughly equally and the balance of technology does not heavily favor the offense statesmay feel secure and signal their peaceful intentions If the balance heavily favors offense howeverstates will face strong incentives to build offensive forces and ght preemptively On the relation-ship between offense-defense theory and structural realism see Van Evera Causes of War pp 7ndash11117 255ndash256 Glaser and Kaufmann ldquoWhat Is the Offense-Defense Balancerdquo pp 48ndash49 Lynn-Jones ldquoOffense-Defense Theory and Its Criticsrdquo pp 660 n 1 664ndash665 and Glaser ldquoRealists AsOptimistsrdquo pp 54 60ndash644 ldquoSelect Enemy Deleterdquo Economist March 8 1997 pp 21ndash24 at p 215 See Jervis ldquoCooperation under the Security Dilemmardquo pp 199ndash201 Glaser and KaufmannldquoWhat Is the Offense-Defense Balancerdquo p 44 and Van Evera ldquoOffense Defense and the Causesof Warrdquo p 40 Other proponents of offense-defense theory are less sanguine about the capacity ofarms control to reduce the likelihood of war pointing out that arms control may be possible onlywhen it is not necessary
































































































































































Conclusions










































































































































































Conclusions




































































































































































Conclusions































































































































































Conclusions


























































































































































Conclusions





















































































































































Conclusions














































































































































Conclusions









































































































































Conclusions




































































































































Conclusions






























































































































Conclusions
























































































































Conclusions



















































































































Conclusions













































































































Conclusions








































































































Conclusions


































































































Conclusions



























































































Conclusions






















































































Conclusions
















































































Conclusions









































































Conclusions



































































Conclusions































































Conclusions


























































Conclusions





















































Conclusions















































Conclusions










































Conclusions




































Conclusions






























Conclusions


























Conclusions






















Conclusions


















Conclusions













Conclusions




















Documents

Grasping the Keir A. Lieber Technological Peace